Anthropic announced it is disabling its Mythos 5 and Claude Fable 5 models following an export control directive from the US government. The order, received on Friday afternoon, asked the company to suspend access for all customers, including foreign nationals and employees, to comply with national security concerns. The move marks a new escalation in tensions between Anthropic and the Trump administration, which has previously labeled the company a supply chain risk.
The company released Claude Fable 5 earlier this week, a version of its Mythos AI model with safeguards to prevent it from answering questions about cybersecurity, biology, and chemistry. Anthropic said it had conducted the public release in collaboration with the US government. The model had a limited rollout in April, aimed at helping companies improve their cybersecurity defenses and address concerns about its potential misuse.
In a blog post, Anthropic stated it received a letter from the US government at 5:21pm ET. The letter did not provide specific details about the national security concern, but the company claimed it understood the government believed it had become aware of a method to bypass, or 'jailbreak,' Fable 5. The company reviewed a demonstration of this technique and identified a small number of previously known, minor vulnerabilities. These vulnerabilities, it said, appear relatively simple and are also found in other publicly available models.
Source: wired
